Looking at this page:
Assuming we only have A input, B do something, C (optional- you'll see why I say this)
Would an iterable B followed by a join C be equivalent to a mapnode B only (no C)?
In other words, the mapnode has a built in join of the outputs as a list?
Related to: https://neurostars.org/t/run-workflow-n-times/76
Eg. Use a mapnode on an inner workflow to have the inner workflow run multiple times?
Unfortunately, a workflow is not an interface, so interface=wf_inner doesn't work in the mapnode.